robot_simulator_rust

A simulator of an autonomous mobile robot which estimates its pose by using Extended Kalman Filter and calculates control input by using Dynamic Window Approach.

description

This program simulates an differential-drive wheeled autonomous robot like below:

• the state of robot: (x, y, θ)
• the control inputs: (v, ω) v=linear velocity, ω=angular velocity

This robot equips a 360-degree camera, and it can observe the distance and angle to the marker on environments.

simulation

circular route

This robot chases a target which moves on a circle.

• target: black object and black line
• actual trajectory: blue line
• observed mark: green lines
• estimated pose: red object
• estimated trajectory: red line

square route

This robot chases a target which moves on a square.

• target: black object and black line
• actual trajectory: blue line
• observed mark: green lines
• estimated pose: red object
• estimated trajectory: red line

waypoints trace

This robot traces a given waypoints.

• target: black object and black line
• actual trajectory: blue line
• observed mark: green lines
• estimated pose: red object
• estimated trajectory: red line

how to use (on macOS Catalina)

start plogger

1. clone robot_simulator (python PJ)

   git clone https://github.com/nmatsui/robot_simulator
   

2. install python 3.9 and dependent libraries by referring to README of robot_simulator

3. start plotter

   pipenv run plotter
   

start agent

1. build a release binary

   cargo build --release
   

2. start the binary with the agent type (circular, square or waypoints)

   ./target/release/robot_simulator_rust circular
   

   or

   ./target/release/robot_simulator_rust square
   

   or

   ./target/release/robot_simulator_rust waypoints
   

background

state-space model

• state-space_model.md

pose estimation using EKF

• ekf.md

local path planning

• local_path_planning.md

license

BSD 3-Clause License

copyright

Copyright (c) 2021, Nobuyuki Matsui